API Trends

API is playing an increasingly critical role in digitization. Postman’s annual “State of API Report” in 2020 reveals that APIs are important to the digital transformation initiatives for 85% of those polled. The enhanced API management expanded 230% year over year, as stated in the “State of API Economy 2021” report by Google/Apigee. 

With tremendous success and momentum in the past few years, we would wonder what’s next for API? Here is the outlook of the trending themes for the API ecosystem to evolve in the digital economy: Standardization, Orchestration, Containerization, Composable, Even-driven, and REST+ (SOCCER).

• Standardization
• Standardization remains the top huddle to grow APIs for 58% organizations, according to the “2020 State of API Survey” conducted by SmartBear. The increased adoption and development of APIs are moving the API design towards more standard-aligned. For example, the OpenAPI Specification (OAS), formerly known as Swagger Specification, has become an industry standard for describing HTTP-based APIs, currently in Version 3. JSON Schema is another specification used for defining the structure of JSON data in a JSON-based format, with the latest version being 2020-12. In the API design, all REST API operations must use the standard HTTP methods, in conformance to HTTP semantics, such as media types. For the sake of loose coupling, Hypermedia as the Engine of Application State (HATEOAS) should be used to allows the request to not only retrieve the data but also get the actions related to the resource. Hypertext Application Language (HAL) is a standard to represent elements on resources and links in JSON and XML. IETF RFC 5988 Web Linking specification also defines the relation types for Web links.
• Orchestration
• The API orchestration is the process of combining multiple applications into a single offering, which commonly merges disparate API calls into a single frontend, like public APIs, partner APIs, internal shared APIs, private APIs, low-level APIs, and composite APIs, It manages interactions, handles caching, facilitate workflows, coordinate entitlements, automates operations, audits traffic, enforces quotas, or aggregates various internal APIs for better user experience. The typical orchestration scenarios include data transfer, field mapping, method combination to workflow, time-based constraints, different API types, and custom scripting. An API gateway or orchestration layer is usually applied for this purpose.
• Containerization
• The vast majority of APIs are implemented as microservices, which are progressively packed with only the operating system libraries and dependencies needed to run the service, resulting in a single lightweight executable. One of the most well-known and widely used container technologies is Docker. The image is more portable and resource-efficient, making APIs self-contained, scalable, and deployable to any cloud or on-prem infrastructure. Cloud-native services are leveraged and serverless is capitalized on as appropriate.
• Composable
• Infrastructure automation, portfolio simplification, shared services, and multicloud are all crucial enablers for composable architecture, which is the process of scaling computing resources in a more dynamic manner. A composable enterprise mixes commercial off the shelf (COTS) products, open-source packages, cloud APIs, SaaS offerings, platform services, and custom development to produce flexible solutions that adapt to fast-changing business demands. The API gateway, development portal, service catalog, lifecycle management, analytics, and API governance services are the main features in the composable APIs. 
• Event-driven
• The event-driven architecture is a type of architecture that processes data in real time in response to state changes or triggers. It decouples systems, allowing microservices-based applications to scale reliably and linearly. More and more organizations are turning to streaming systems like Kafka for pub/sub. AsyncAPI is utilized as a standard to describe event-driven microservices to cover the entire spectrum: discovery, specification, channel, server, security, code generation, and event managing. AsyncAPI specifies the technology-agnostic interface of event-driven systems, regardless of the underlying implementation. For instance, a webhook is a user-defined HTTP callback, used  as a technique for custom callbacks triggered by some events to augment or modify the functionality of a web page or online application. MQTT is another event-driven messaging example, which a lightweight highly-scalable pub/sub protocol that transports messages between devices and clouds, widely used in IoT.
• REST+
• RESTful web services are dominant API implementations. Beyond REST, GraphQL provides a comprehensive and comprehensible explanation of the data in the API, allowing callers to request only what they require. GraphQL is leveraged as a query language for APIs and a runtime for executing queries. gRPC is another open-source framework for making remote procedure calls, which employs HTTP/2 for transport and Protocol Buffers for interface description, with authentication, blocking or nonblocking bindings, bidirectional streaming, flow control, cancellation, and timeouts. On the other hand, a WebSocket provides a full-duplex communication channel over a single connection between a client and server. In addition, WebSub is exploited as an open protocol for a large amount of data in distributed pub/sub communications between publishers, subscribers, and hubs over the Internet.

It is foreseen that broader adoptions and innovations will take place in these six areas. APIs are expected to catalyze the accelerated digital transformations in all organizations in the near future.

For more information, please contact Tony Shan (blog@tonyshan.com) or leave your comments below.

   ©Tony Shan. All rights reserved. All standard disclaimers apply here.